Potential and limitations of machine-learning approaches to inclusive $|V_{ub}|$ determinations

TL;DR

This paper evaluates machine learning methods, particularly deep neural networks, for improving the inclusive determination of |V_{ub}| in B meson decays, highlighting their potential and limitations in background separation and kinematic coverage.

Contribution

It compares low-level neural network features with high-level physicist-engineered features for |V_{ub}| extraction, revealing trade-offs in signal acceptance and kinematic dependence.

Findings

Deep neural networks offer modest background separation improvements.

Neural networks excluding kinematic features are flatter but less inclusive.

Trade-offs depend on Monte Carlo models like Sherpa and EvtGen.

Abstract

The determination of $|V_{ub}|$ in inclusive semileptonic $B \to X_u \ell \nu$ decays will be among the pivotal tasks of Belle II. In this paper we study the potential and limitations of machine learning approaches that attempt to reduce theory uncertainties by extending the experimentally accessible fiducial region of the $B\to X_u \ell\nu$ signal into regions where the $B\to X_c \ell\nu$ background is dominant. We find that a deep neural network trained on low-level single particle features offers modest improvement in separating signal from background, compared to BDT set-ups using physicist-engineered high-level features. We further illustrate that while the signal acceptance of such a deep neural network deteriorates in kinematic regions where the signal is small, such as at high hadronic invariant mass, neural networks which exclude kinematic features are flatter in kinematics but less inclusive in the sampling of exclusive hadronic final states at fixed kinematics. The trade-off between these two set-ups is somewhat Monte Carlo dependent, and we study this issue using the multipurpose event generator Sherpa in addition to the widely used $B$-physics tool EvtGen.